If you’re having trouble getting your Agent to connect to your Oracle Database via our SQLcl MCP Server, it might be due to a missing TNS_ADMIN environment variable. Here’s how to fix that!
I’ll be joining aioug and the Oracle AI World Tour in Feburary 2026. This will be my first trip to India, so I’m very excited to see this amazing part of the world. Details I’ll have talks on getting the best Agentic AI experience from your Oralce AI Database and an overview of our database tooling for Oracle Cloud Infrastructure. You can of course also expect lots of tips & tricks, and some sneak peeks…
Disclaimer: I have a hard time ‘unplugging’ from work. This post isn’t a humble-brag. I’m frequently asked what AI is ‘good for’ when it comes to databases. And since we have an MCP Server for Oracle AI Database, then I have some pretty good ideas on that. Yesterday I heard this on SQL optimization – And yes, it can be really good at SQL type tasks. But it’s more than just making ‘sql faster,’ its…
With the release of Oracle SQLcl version 25.4, your MCP Server can now accept AI Agent requests to run SQL statements and scripts in the background, and request the results… later.
Question: can we build our own AI Agent and use your MCP Server to interact with our Oracle AI Database? Answer: Yes! See Anders’ Java example, it’s awesome!
After a very long break, I’ve released two new YouTube channel videos over the past 2 days! The first is an overview of our MCP Server, & the 2nd is all demo.
Use Oracle SQL Developer extension for VS Code and our MCP Server to annotate your data & assist your agents for more efficient NL2SQL interactions.
Let’s build a fancy web report with AI and our MCP Server for Oracle Database, complete with a search feature that uses 26ai Vector and ORDS REST APIs.
Version 25.3 of our SQL Developer Extension for VS Code is now available, featuring couple of really nice new features and a TON of bug fixes.
I asked Claude to dev a REST API for my databse. But I didn’t like the SQL. So I used my brains to fix it, update the API, & then test/compare the performance.